Open Ocean Dead-Zones in the Tropical J I Northeast Atlantic

Home , Particle (ecology)

icsinPpr|Dsuso ae icsinPpr|Dsuso ae | Paper Discussion | Paper Discussion | Paper Discussion | Paper Discussion Biogeosciences Discuss., 11, 17391–17411, 2014 www.biogeosciences-discuss.net/11/17391/2014/ doi:10.5194/bgd-11-17391-2014 BGD © Author(s) 2014. CC Attribution 3.0 License. 11, 17391–17411, 2014

This discussion paper is/has been under review for the journal Biogeosciences (BG). Open ocean Please refer to the corresponding ﬁnal paper in BG if available. dead-zones Open ocean dead-zone in the tropical J. Karstensen et al. North Atlantic Ocean Title Page 1 1 1 1 1 2 J. Karstensen , B. Fiedler , F. Schütte , P. Brandt , A. Körtzinger , G. Fischer , Abstract Introduction R. Zantopp1, J. Hahn1, M. Visbeck1, and D. Wallace3 Conclusions References 1GEOMAR Helmholtz Centre for Ocean Research Kiel, Kiel, Germany Tables Figures 2Faculty of Geosciences and MARUM, University of Bremen, Bremen, Germany 3Halifax Marine Research Institute (HMRI), Halifax, Canada J I Received: 3 November 2014 – Accepted: 13 November 2014 – Published: 12 December 2014 J I Correspondence to: J. Karstensen ([email protected]) Back Close Published by Copernicus Publications on behalf of the European Geosciences Union. Full Screen / Esc

Printer-friendly Version

Interactive Discussion

17391 icsinPpr|Dsuso ae icsinPpr|Dsuso ae | Paper Discussion | Paper Discussion | Paper Discussion | Paper Discussion Abstract BGD The intermittent appearances of low oxygen environments are a particular thread for marine ecosystems. Here we present ﬁrst observations of unexpected low (< 11, 17391–17411, 2014 2 µmolkg−1) oxygen environments in the open waters of the eastern tropical North −1 5 Atlantic, a region where typically oxygen concentration does not fall below 40 µmolkg . Open ocean The low oxygen zones are created just below the mixed-layer, in the euphotic zone of dead-zones high productive cyclonic and anticyclonic-modewater eddies. A dynamic boundary is created from the large swirl-velocity against the weak background ﬂow. Hydrographic J. Karstensen et al. properties within the eddies are kept constant over periods of several months, while net 10 respiration is elevated by a factor of 3 to 5 reducing the oxygen content. We repeatedly observed low oxygen eddies in the region. The direct impact on the ecosystem is Title Page evident from anomalous backscatter behaviour. Satellite derived global eddy statistics Abstract Introduction do not allow to estimate the large-scale impact of the eddies because their vertical structure (mixed-layer depth, euphotic depth) play a key role in creating the low oxygen Conclusions References 15 environment. Tables Figures

1 Introduction J I J I The concentration of dissolved oxygen (DO) in seawater is of critical importance to almost all marine life and oceanic biogeochemical cycling (Wilding, 1939; Diaz and Back Close Rosenberg, 2008; Vaquer-Sunyer and Duarte, 2008; Wright et al., 2012; Kalvelage Full Screen / Esc 20 et al., 2011). Local DO concentrations are the result of a delicate balance between

oxygen supply and consumption and eventually regions of extreme low oxygen Printer-friendly Version content are created: at the microscale at particle boundaries (Alldredge and Cohen, 1987), at the mesoscale as coastal dead-zones (Diaz and Rosenberg, 2008) or at Interactive Discussion the large scale such as eastern boundary Oxygen Minimum Zones (OMZ) (Luyten 25 et al., 1983; Karstensen et al., 2008). Understanding processes that control the

17392 icsinPpr|Dsuso ae icsinPpr|Dsuso ae | Paper Discussion | Paper Discussion | Paper Discussion | Paper Discussion DO supply/consumption balance, and any possible alterations over time, remain a challenge of current research. BGD Critical DO concentration thresholds have been identified, which lead to major 11, 17391–17411, 2014 reorganizations of the marine ecosystems (Vaquer-Sunyer and Duarte, 2008; Wright 5 et al., 2012; Kalvelage et al., 2011). For higher trophic levels, such as fish, the impact of a certain DO levels on metabolism and as such fitness, is species dependent Open ocean (Wilding, 1939). Nevertheless, for DO below 20 µmolkg−1 (“severe hypoxia”) mass dead-zones mortality of fish has been reported (Diaz and Rosenberg, 2008). Also triggered by severe hypoxia DO levels, microbes begin to convert nitrite and ammonium to nitrogen J. Karstensen et al. 10 gas, and this remove fixed nitrogen from the water which in turn limits primary productivity (Wright et al., 2012; Kalvelage et al., 2011). A next distinct DO threshold Title Page is for concentrations below about 5 µmolkg−1 when microbes begin to utilize nitrate (and other nitrogen species) as terminal electron acceptors in anaerobic respiration Abstract Introduction (“denitrification”) (Wright et al., 2012; Kalvelage et al., 2011). Finally, when DO reach −1 Conclusions References 15 concentrations around 1 µmolkg (“anoxia”), only specifically adapted microbes can exist (Wright et al., 2012). Tables Figures The pelagic zones of the eastern tropical North Atlantic OMZ are considered to −1 be “hypoxic”, with minimal DO of hardly below 40 µmolkg (Stramma et al., 2009; J I Karstensen et al., 2008). As such it is assumed that the DO levels pose on the regional J I 20 ecosystem some limitation in biodiversity primarily through avoidance and maybe an increased mortality (Vaquer-Sunyer and Duarte, 2008). The region is thus very much Back Close in contrast to the major OMZs in the eastern North and South Pacific Ocean and Full Screen / Esc the northern Indian Ocean where DO concentrations pass all DO thresholds outlined above, and as such specifically adapted ecosystems must exists. Printer-friendly Version 25 However, here we report on recent observations of extreme low DO, characterizing severe hypoxia and even anoxia with horizontal scales of about 100 km and vertical Interactive Discussion scales of about 100 m, exists in the eastern tropical North Atlantic.

17393 icsinPpr|Dsuso ae icsinPpr|Dsuso ae | Paper Discussion | Paper Discussion | Paper Discussion | Paper Discussion 2 Data and methods BGD 2.1 Moored oxygen sensors 11, 17391–17411, 2014 DO time series were acquired at the Cape Verde Ocean Observatory (CVOO) mooring. CVOO is located in the eastern tropical North Atlantic, about 100 km northeast of the ◦ 0 ◦ 0 Open ocean 5 Cape Verde Island Sao Vicente (17 35 N, 24 15 W) and approximately 800 km from dead-zones the Mauritanian coast (Fig.1). Since 2006 the observatory is equipped with oxygen optodes, while since the beginning of 2008 at least one optode was installed in the J. Karstensen et al. upper 60 m of the water colmn. Oxygen measurements at the CVOO mooring were done with AADI Aanderaa 10 oxygen optodes (type 3830). A calibration was done against CTD/oxygen sensor Title Page (Sea-Bird Electronics 43 Clark electrode) data for the deployment period from 2009 Abstract Introduction to 2014 before and after the deployment phase. For each optode, 17 independent calibration points were obtained, choosing locations with weak vertical oxygen gradient. Conclusions References A lab calibration at zero oxygen concentration was also performed by submerging the Tables Figures 15 optodes into a sodium sulfite solution. All calibration points were used to derive a final calibration equation, while the chemically forced (and thus more precise) zero oxygen J I calibration was weighted three times higher than the CTD/oxygen cast references. The difference between calibration point observations and calibrated optode suggests an J I −1 overall rms error of 3 µmolkg . Comparison of the chemically forced zero oxygen Back Close 20 phase data and the phase readings during the anoxic event suggests a higher accuracy of about 1 µmolkg−1. Pressure and salinity variability was corrected according to the Full Screen / Esc AADI manual. Printer-friendly Version 2.2 Argo float Interactive Discussion Data from a PROVOR Profiling Argo Float (WMO 6900632; Martec Inc., France), 25 equipped with a standard CTD (SBE 41CP; Sea-Bird Electronics, Bellevue, USA), an oxygen optode (3830; Aanderaa Data Instruments, Bergen, Norway) and 17394 icsinPpr|Dsuso ae icsinPpr|Dsuso ae | Paper Discussion | Paper Discussion | Paper Discussion | Paper Discussion a transmissometer (CRV5; WETLabs, Philomath, USA) was used. The float was programmed to conduct every 5th day a vertical profile between 400 db and the BGD surface with a vertical resolution of 5 db. Transmitted DO concentrations were corrected 11, 17391–17411, 2014 for salinity effects and a pressure correction of 4 % per 1000 db was applied to 5 the data. After deployment, a ship-borne CTD cast that included measurements for DO (Winkler Titration and a Sea-Bird SBE 43 Clark Electrode), was conducted next Open ocean to the deployment location and used for post offset correction of float-based DO dead-zones measurements. Transmissometer data was obtained in units of m−1 (beam attenuation coefficient) based on the factory calibration. Biofouling within the optical path was J. Karstensen et al. 10 accounted for by subsequently subtracting each profile’s minimum beam attenuation value from the respective profile data. Title Page

2.3 Eddy tracking Abstract Introduction

For detecting and tracking eddy-like structures from the daily merged SLA data, an Conclusions References

algorithm based on the Okubo–Weiß method was applied. The method is robust and Tables Figures 15 widely used to detect eddies in satellite data as well as on numerical model output (Chelton et al., 2007; Sangrà et al., 2009). Analysing the contribution of relative vorticity J I on the strain tensor, an eddy is deﬁned as a region of negative W (vorticity dominates over strain) surrounded by a region of positive W (strain dominates over vorticity). We J I W −12 −2 applied a threshold of 0 = −2 × 10 s as our eddy detection limit. Moreover, we Back Close 20 considered that the detected area has a circular area-equivalent with a radius larger than 40 km and must be detectable for at least 7 days. An eddy trajectory is derived by Full Screen / Esc following the centre of individual W0 areas in SLA maps from 1 day (maximum 10 km) up to 3 weeks (maximum distance 60 km). Printer-friendly Version

Interactive Discussion

17395 icsinPpr|Dsuso ae icsinPpr|Dsuso ae | Paper Discussion | Paper Discussion | Paper Discussion | Paper Discussion 3 Results BGD 3.1 Open ocean low oxygen events from moored observation 11, 17391–17411, 2014 The ﬁrst data sets we report is DO time series from CVOO mooring available from 2006 until 2014 (Fig.2). The time series from the instruments installed shallower than −1 Open ocean 5 60 m depth typically shows short term variability of 50 µmolkg around the saturation. dead-zones However, exceptional low oxygen events are observed during boreal winter 2007, 2010, 2011, and 2012. J. Karstensen et al. The most intense event was recorded at the mooring site over a period of about 1 month, in February 2010 (CVOO2010; Fig.3a) with DO concentrations close to −1 10 zero (DO < 2 µmolkg ) at shallow depth (42 m optode), where typical DO is within Title Page 10 % of saturation (> 200 µmolkg−1). At a second sensor, installed at nominal 170 m, Abstract Introduction showed a more moderate DO decreased from the typical 100 µmolkg−1 to less than 30 µmolkg−1 during the event. Inspecting the hydrography and currents from other Conclusions References moored instruments we found that the low DO event was accompanied by a lens of Tables Figures 15 cold and less saline (Fig.3b) water and a strong and reversing meridional ﬂow (Fig.3c). The temporal evolution of the isopycnals (surfaces of constant water density) indicates J I that an anticyclonic-modewater eddy has passed through the mooring and contained the extreme low DO concentrations. J I

Anticyclonic-modewater eddies, also called intrathermocline eddies, are associated Back Close 20 with downward/upward bended isopycnals below/above a subsurface velocity maximum (Kostianoy and Belkin, 1989). The transition between up and downward Full Screen / Esc bended isopycnals form a lens (or mode) of a speciﬁc water mass. In our observations the mode has a height of about 50 m and is also centred at about 50 m depth Printer-friendly Version

and, as a consequence, the surface mixed-layer shoaled from a thickness of about Interactive Discussion 25 50–60 m before (and after) the eddy passage to less than 20 m during the passage. Anticyclonic-modewater eddies have been reported before to be associated with speciﬁc ecosystem responses, such as high primary productivity presumably due to

17396 icsinPpr|Dsuso ae icsinPpr|Dsuso ae | Paper Discussion | Paper Discussion | Paper Discussion | Paper Discussion eddy-wind interaction (McGillicuddy et al., 2007) while controversy about the governing processes exists (Mahadevan et al., 2008). BGD Earlier DO record from CVOO revealed that in February 2007 (CVOO2007), 11, 17391–17411, 2014 almost exactly 3 years before the 2010 event, a single oxygen sensor installed at 5 120 m (nominal) depth also recorded very low DO concentrations, down to about 15 µmolkg−1. This low oxygen event was again associated with the passage of an Open ocean anticyclonic-modewater eddy. dead-zones

3.2 Open ocean low oxygen events in Argo ﬂoat data J. Karstensen et al.

One further North Atlantic “severe hypoxia” event was detected within a cyclonic 10 eddy and surveyed with an Argo type float (Fig.4a and b) that operated from mid Title Page February 2008 until the end of May 2009. After launch, the float remained in the Abstract Introduction Mauritanian upwelling region until, by the end of May 2008, it begun to move in west- northwest directions (Fig.1). The float trajectory already indicated that it was trapped in Conclusions References

a cyclonic eddy, as conﬁrmed by a more detailed analysis presented below. In parallel Tables Figures 15 to the westward propagation into the open waters a decrease in DO was seen at all depth and that lasted until mid-December 2008. During that period, lowest DO (about J I 14 µmolkg−1) was always observed close to the mixed-layer base which, however, successively deepened. After December 2008 the DO increased again (Fig.4b), J I

accompanied by drastic changes in temperature (not shown here) and salinity (Fig.4a). Back Close 20 It turned out from the eddy trajectory analysis below, that the float was still inside the eddy but the changes in the water properties indicate that the eddy lost its isolated Full Screen / Esc character. Printer-friendly Version 3.3 Propagation of oxygen anomalies Interactive Discussion Making use of sea-level anomaly (SLA) maps, all three low oxygen observations 25 (CVOO2007, Argo2008, CVOO2010) could be associated with individual nonlinear mesoscale eddies (Chelton et al., 2011). The two eddies observed at CVOO rotated 17397 icsinPpr|Dsuso ae icsinPpr|Dsuso ae | Paper Discussion | Paper Discussion | Paper Discussion | Paper Discussion in an anticyclonic way (as also seen in the CVOO mooring data), the Argo2008 eddy was a cyclonic eddy. Considering the along path characteristics from concurrent SLA BGD maps, all three eddies had roughly similar diameters (about 130 km) and propagated 11, 17391–17411, 2014 westward with speed of about 7 cm s−1. As such they can be categorized as “typical” 5 for this latitude range (Chelton et al., 2011; Chaigneau et al., 2009). The SLA across the eddy radius was rather weak with an amplitude of 1.5 (±1.5) cm (negative for Open ocean the cyclone, positive for the anticyclones). The SLA anomaly translates to maximum dead-zones geostrophic surface currents of about 0.05–0.10 m s−1, which is slow when compared with global eddies characteristics (Chelton et al., 2011; Risien and Chelton, 2008). J. Karstensen et al. 10 Moreover, such a slow surface geostrophic current reflect only 10–20 % of the oberserved interior maximal swirl velocity, at least for the anticyclonic-modewater Title Page eddies (Fig.3c). All three eddies had a similar region of origin at about 18 ◦ N/16.5◦ W. The cyclone was formed in May 2008 (the float entered the eddy core about one month Abstract Introduction later), and the two anticyclonic-modewater eddies in July 2006 and 2009, respectively. Conclusions References 15 Cyclonic eddies as well as anticyclonic-modewater eddies have been shown to be vital for the oceanic productivity (Chelton et al., 2011; McGillicuddy et al., 2007). Cloud Tables Figures free/partial cloud conditions are rare in the region, but selected satellite derived surface chlorophyll images clearly show such productivity events (Fig.5), covering almost the J I whole diameter of the eddies. J I

20 3.4 Isolation and respiration Back Close

In July 2006 the research vessel METEOR operated close to the coast and surveyed Full Screen / Esc by chance the CVOO2007 eddy, shortly after it left the Mauritanian upwelling region (Fig.6). Comparing the ships survey data from 2006 with the mooring survey data Printer-friendly Version

from 2007 reveals that the core of the eddy remained unchanged in temperature Interactive Discussion 25 and salinity over a period of seven months, and after propagating more than 650 km westward. Moreover, the water within the eddy shows distinct South Atlantic Central Water characteristics, typical found in the Mauritanian upwelling region.

17398 icsinPpr|Dsuso ae icsinPpr|Dsuso ae | Paper Discussion | Paper Discussion | Paper Discussion | Paper Discussion For the anticyclonic-modewater eddies (CVOO2007, CVOO2010) the direct velocity observations show that just below the mixed-layer base the ratio (α) of swirl velocity BGD to translation speed, which is a proxy for the nonlinearity (and coherence) of eddies 11, 17391–17411, 2014 (Chelton et al., 2011), is at maximum (α > 9). Note, at this depth not only highest 5 swirl velocities but minimal DO concentrations are found. No direct swirl velocity observation exists for cyclone surveyed by the Argo2008 but composite density profiles, Open ocean recorded when the float entered the eddy (May/June 2008) and when it left the eddy dead-zones (March/April 2009), were used to derive geostrophic current profiles (see Chaigneau et al., 2011). As we do not have a proper reference velocity the absolute value for the J. Karstensen et al. 10 α, and as such for the intensity of the isolation, cannot be calculated. However, we observe a fundamental change in the velocity shear profile, from a rotation with nearly Title Page constant velocity from just below the mixed-layer (30 m) at the beginning of the survey to a profile with a distinct peak in swirl velocity below 100 m depth at the end of the Abstract Introduction float survey. Such a change in the flow structure indicate that the maximum α moved Conclusions References 15 to deeper levels, and the eddy lost its isolated character at shallow depth, which in turn may explain the strong change in hydrographic properties (and the oxygen increase) Tables Figures as lateral exchange with surrounding waters could take place at this stage of the eddies life-time (Fig.4). J I From differences in DO concentrations between concurrent Argo float profiles, an J I 20 apparent oxygen utilization rate (aOUR) profile was estimated (Fig.4c). Only profiles where DO decreased (May to mid-December 2008) were considered (isolated eddy). Back Close In order to take the successive deepening of isopycnals over time into account (e.g. Full Screen / Esc Fig.4a, contours) the aOUR was calculated in density classes and subsequently projected back to depth, using the mean vertical density profile. The aOUR is highest, −1 −1 Printer-friendly Version 25 more than 0.15 µmolkg day , just below the mixed-layer and level out to about −1 −1 0.05 µmolkg day between 120 m and the maximum depth the float surveyed Interactive Discussion (400 m). These rates are 3 to 5 times higher than typical rates for the thermocline (Jenkins, 1982; Karstensen et al., 2008) but as the calculations do not consider any lateral oxygen supply (which should be minimal given the constancy of the hydrographic

17399 icsinPpr|Dsuso ae icsinPpr|Dsuso ae | Paper Discussion | Paper Discussion | Paper Discussion | Paper Discussion properties), the rates can be seen as lower bounds for the respiration inside of the eddy. The relative particle load, as observed with the transmissometer (not shown), is BGD at maximum just at the base of the mixed-layer. However, the minimal DO is observed 11, 17391–17411, 2014 about 5 m below that particle maximum and which indicates that the net oxygen 5 respiration is related to the sinking particles. For the anticyclonic-modewater eddies a net DO respiration can only be derived Open ocean for the CVOO2007 eddy and for one depth only. This eddy was surveyed twice, oﬀ dead-zones Mauritania by RV METEOR and 7 month later at the CVOO mooring site. Between these two surveys the DO concentrations at the 120 m depth level (only depth with J. Karstensen et al. −1 10 DO instrument at CVOO2007) changed by more than 50 µmolkg . Such a change translate to an aOUR of 0.25 µmolkg−1 day−1 and which is even higher than in the Title Page cyclonic eddy and suggests that the productivity (and subsequently oxygen drawdown through sinking particles) is more intense in anticyclonic-modewater eddies then in Abstract Introduction cyclonic eddies. Conclusions References

Tables Figures 15 4 Discussion

Here we have reported first observations of open ocean dead-zones in the tropical J I Northeast Atlantic. We could show their evolution, from their generation near the J I Northwest African shelf to the westward propagation into the open ocean during which Back Close extrem low-oxygen environments developed just below the mixed layer. To estimate the 20 impact of the eddies on the large-scale ocean at least their intensity and occurrence Full Screen / Esc must be known. While methods for eddy surface detection from SLA maps are rather well established (Chelton et al., 2007, 2011; Chaigneau et al., 2009), concurrent details Printer-friendly Version on the vertical stratification are required to detect the dead-zone eddies. A possibility Interactive Discussion has been outlined using Argo float data in parallel to SLA data (Chaigneau et al., 2009) 25 to map the large-scale mean hydrographic eddy (cyclone, anticyclone) structures. Nevertheless, for a dead-zone detection the DO cencentration is required as well and which is so far not a routinely measured with floats. 17400 icsinPpr|Dsuso ae icsinPpr|Dsuso ae | Paper Discussion | Paper Discussion | Paper Discussion | Paper Discussion There is clearly a local impact of dead-zone eddies on the ecosystem. During the passage of the anticyclonic-modewater eddies at the CVOO we observed in the target BGD strength data that acoustic back scatterers, such as zooplankter, stopped their diurnal 11, 17391–17411, 2014 migration cycle (Fig. 7). Such an absence of the vertical migration is indicative for 5 zooplankter in the major OMZ regions (Ayon et al., 2008). While in the open ocean mobile organisms may escape from the dead-zone other organisms, such as the wide Open ocean range of prokaryotes, may need to adapt to the environment in order to survive. In dead-zones that sense the dead-zone eddies can be seen as gigantic natural laboratories where an extreme environment is created in a relative short period of time (a few month). As J. Karstensen et al. 10 such these features may open new way in investigating the adaptation techniques of organisms to low DO environments. Title Page It is worth considering the possible interaction of a dead-zone eddy with an island (e.g. Cape Verdes Islands archipelago). Given the shallow depth the low DO is located, Abstract Introduction a sudden flooding of a coastal areas with low DO waters will have a dramatic impact on Conclusions References 15 the local ecosystems, and sudden fish or crustacean death may occur. In retrospective, such eddy/island interactions may explain such events that have been reported in the Tables Figures past (O. Melicio, personal communication, National Fisheries Institute INDP, Mindelo, Sao Vicente, Cape Verde Island). J I

J I

5 Conclusions Back Close

20 Differences in the eddy dynamics among cyclones and anticyclonic-modewater eddies Full Screen / Esc can explain differences between the intensity and lifetime of their respective dead- zones. For example the generation of up- and downwelling within eddies through Printer-friendly Version interaction with the overlying wind field (McGillicuddy et al., 2007; Mahadevan et al., Interactive Discussion 2008) can generate anomalous shallow (deep) mixed-layers in the anticyclonic- 25 modewater (cyclonic) eddies which in turn may influence productivity, export and respiration efficiency (Klein and Lapeyre, 2009; Lévy et al., 2012). Further observational and modeling studies are required that focus on the submesoscale 17401 icsinPpr|Dsuso ae icsinPpr|Dsuso ae | Paper Discussion | Paper Discussion | Paper Discussion | Paper Discussion dynamics and on the biogeochemical cycling and ecosystem response within the eddies and their potential impact on large-scale marine ecosystem (Klein and Lapeyre, BGD 2009; Lévy et al., 2012). 11, 17391–17411, 2014 In principle open ocean dead-zones in mesoscale eddies may be created in all 5 oceanic regions characterized through a sluggish background flow. Primary sites are the eastern boundary shadow zones (Luyten et al., 1983). Given the much lower Open ocean oxygen levels of the major OMZs in the Pacific or the Indian Ocean, the dead-zones in dead-zones those regions will not appear in DO anomalies but in other biogeochemical extremes, such as an anomalous nitrogen isotope composition (Altabet et al., 2012). J. Karstensen et al. 10 One may wonder why dead-zone eddies have not been discovered before? Besides the undersampling of the tropical eastern North Atlantic in the past, there is the Title Page likelihood that such low oxygen concentrations were disregarded as “outliers” in the data sets. In fact we interpreted the low oxygen at CVOO2007 first as an outlier related Abstract Introduction to an instrumental error and only the more recent events recorded with the double Conclusions References 15 sensor package at CVOO2010, and combined with sophisticated optode calibration procedures, gave us confidence that our observations were real. Tables Figures

Acknowledgements. Financial support is acknowledged from: the European Commission for FP7 GROOM (284321), FP7 CARBOOCEAN (264879), FP7 CARBOCHANGE (264879), FP7 J I FixO3 (312463); the DFG for Collaborative Research Centre “SFB 754”; the BMBF for SOPRAN J I 20 (03F0462A) and AWA (01DG12073E). CVOO is part of the OceanSITES network. SLA data was produced by Ssalto/Duacs and distributed by Aviso (www.aviso.oceanobs.com). MyOcean Back Close products have been used. Argo ﬂoat data is available from Coriolis data centre (www.coriolis. Full Screen / Esc eu.org).

Printer-friendly Version References Interactive Discussion

25 Alldredge, A. L. and Cohen, Y.: Can microscale chemical patches persist in the sea? Microelectrode study of marine snow, fecal pellets, Science, 235, 689–691, doi:10.1126/science.235.4789.689, 1987. 17392

17402 icsinPpr|Dsuso ae icsinPpr|Dsuso ae | Paper Discussion | Paper Discussion | Paper Discussion | Paper Discussion Altabet, M. A., Ryabenko, E., Stramma, L., Wallace, D. W. R., Frank, M., Grasse, P., and Lavik, G.: An eddy-stimulated hotspot for fixed nitrogen-loss from the Peru oxygen minimum BGD zone, Biogeosciences, 9, 4897–4908, doi:10.5194/bg-9-4897-2012, 2012. 17402 Ayon, P., Criales-Hernandez, M. I., Schwamborn, R., and Hirche, H.-J.: Zooplankton research 11, 17391–17411, 2014 5 off Peru: a review, Prog. Oceanog., 79, 238–255, doi:10.1016/j.pocean.2008.10.020, 2008. 17401 Chaigneau, A., Eldin, G., and Dewitte, B.: Eddy activity in the four major up- Open ocean welling systems from satellite altimetry (1992–2007), Prog. Oceanogr., 83, 117–123, dead-zones doi:10.1016/j.pocean.2009.07.012, 2009. 17398, 17400 J. Karstensen et al. 10 Chaigneau, A., Le Texier, M., Eldin, G., Grados, C., and Pizarro, O.: Vertical structure of mesoscale eddies in the eastern South Pacific Ocean: a composite analysis from altimetry and Argo profiling floats, J. Geophys. Res., 116, C11025, doi:10.1029/2011JC007134, 2011. 17399 Title Page Chelton, D., Schlax, M., Samelson, R., and de Szoeke, R. A.: Global observations of large Abstract Introduction 15 oceanic eddies, Geophys. Res. Lett., 34, L15606, doi:10.1029/2007GL030812, 2007. 17395, 17400 Conclusions References Chelton, D. B., Schlax, M. G., and Samelson, R. M.: Global observations of nonlinear mesoscale eddies, Prog. Oceanogr., 91, 167–216, doi:10.1016/j.pocean.2011.01.002, 2011. Tables Figures 17397, 17398, 17399, 17400 20 Diaz, R. J. and Rosenberg, R.: Spreading dead zones and consequences for marine J I ecosystems, Science, 321, 926–929, doi:10.1126/science.1156401, 2008. 17392, 17393 Jenkins, W. J.: Oxygen utilization rates in North Atlantic subtropical gyre and primary production J I in oligotrophic systems, Nature, 300, 246–248, doi:10.1038/300246a0, 1982. 17399 Kalvelage, T., Jensen, M. M., Contreras, S., Revsbech, N. P., Lam, P., Günter, M., Back Close 25 LaRoche, J., Lavik, G., and Kuypers, M. M. M.: Oxygen sensitivity of anammox Full Screen / Esc and coupled N-cycle processes in oxygen minimum zones, PLoS ONE, 6, e29299, doi:10.1371/journal.pone.0029299, 2011. 17392, 17393 Karstensen, J., Stramma, L., and Visbeck, M.: Oxygen minimum zones in the Printer-friendly Version eastern tropical Atlantic and Pacific oceans, Prog. Oceanogr., 77, 331–350, Interactive Discussion 30 doi:10.1016/j.pocean.2007.05.009, 2008. 17392, 17393, 17399 Klein, P. and Lapeyre, G.: The oceanic vertical pump induced by mesoscale and submesoscale turbulence, Annual Review of Marine Science, 1, 351–375, doi:10.1146/annurev.marine.010908.163704, 2009. 17401, 17402 17403 icsinPpr|Dsuso ae icsinPpr|Dsuso ae | Paper Discussion | Paper Discussion | Paper Discussion | Paper Discussion Kostianoy, A. and Belkin, I.: A survey of observations on intrathermocline eddies in the world ocean, in: Mesoscale/Synoptic Coherent Structures in Geophysical Turbulence, edited by: BGD Nihoul, J. and Jamart, B., vol. 50, Elsevier, New York, 821–841, 1989. 17396 Lévy, M., Ferrari, R., Franks, P. J. S., Martin, A. P., and Rivière, P.: Bringing physics to life at the 11, 17391–17411, 2014 5 submesoscale, Geophys. Res. Lett., 39, L14602, doi:10.1029/2012GL052756, 2012. 17401, 17402 Luyten, J., Pedlosky, J., and Stommel, H.: The ventilated thermocline, J. Phys. Oceanogr., 13, Open ocean 292–309, 1983. 17392, 17402 dead-zones Mahadevan, A., Thomas, L. N., and Tandon, A.: Comment on “Eddy/wind inter- J. Karstensen et al. 10 actions stimulate extraordinary mid-ocean plankton blooms”, Science, 320, p. 448 doi:10.1126/science.1152111, 2008. 17397, 17401 McGillicuddy, D. J., Anderson, L. A., Bates, N. R., Bibby, T., Buesseler, K. O., Carlson, C. A., Davis, C. S., Ewart, C., Falkowski, P. G., Goldthwait, S. A., Hansell, D. A., Jenkins, W. J., Title Page Johnson, R., Kosnyrev, V. K., Ledwell, J. R., Li, Q. P., Siegel, D. A., and Steinberg, D. K.: Abstract Introduction 15 Eddy/wind interactions stimulate extraordinary mid-ocean plankton blooms, Science, 316, 1021–1026, doi:10.1126/science.1136256, 2007. 17397, 17398, 17401 Conclusions References Risien, C. M. and Chelton, D. B.: A global climatology of surface wind and wind stress fields from eight years of QuikSCAT scatterometer data, J. Phys. Oceanogr., 38, 2379–2413, Tables Figures doi:10.1175/2008JPO3881.1, 2008. 17398 20 Sangrà, P., Pascual, A., Ángel Rodríguez-Santana, Machín, F., Mason, E., McWilliams, J. C., J I Pelegrí, J. L., Dong, C., Rubio, A., Arístegui, J., Ángeles Marrero-Díaz, Hernández- Guerra, A., Martínez-Marrero, A., and Auladell, M.: The Canary Eddy Corridor: a major J I pathway for long-lived eddies in the subtropical North Atlantic, Deep-Sea Res. Pt. I, 56, 2100–2114, 2009. 17395 Back Close 25 Stramma, L., Visbeck, M., Brandt, P., Tanhua, T., and Wallace, D.: Deoxygenation in the oxygen Full Screen / Esc minimum zone of the eastern tropical North Atlantic, Geophys. Res. Lett., 36, L20607, doi:10.1029/2009GL039593, 2009. 17393 Vaquer-Sunyer, R. and Duarte, C. M.: Thresholds of hypoxia for marine biodiversity, P. Natl. Printer-friendly Version Acad. Sci. USA, 105, 15452–15457, doi:10.1073/pnas.0803833105, 2008. 17392, 17393 Interactive Discussion 30 Wilding, J.: The oxygen threshold for three species of fish, Ecology, 20, 253–263, 1939. 17392, 17393 Wright, J. J., Konwar, K. M., and Hallam, S. J.: Microbial ecology of expanding oxygen minimum zones, Nature Rev. Microbiol., 10, 381–394, doi:10.1038/nrmicro2778, 2012. 17392, 17393 17404 icsinPpr|Dsuso ae icsinPpr|Dsuso ae | Paper Discussion | Paper Discussion | Paper Discussion | Paper Discussion BGD 11, 17391–17411, 2014

25oN Open ocean

23oN dead-zones

J. Karstensen et al. 21oN 01.05.09

19oN 30.06.07 Title Page 30.06.09 R/V Meteor o CVOO 01.05.08 Abstract Introduction 17 N 03.11.10 (19.07.2006) 01.07.06

Conclusions References 15oN Tables Figures

13oN 38oW 34oW 30oW 26oW 22oW 18oW 14oW J I

Figure 1. Overview map of the eastern tropical North Atlantic Oxygen Minimum Zone. The J I tracks of the two anticyclonic-modewater eddies (CVOO2007, grey line and CVOO2010, black line) observed at CVOO (plus sign; 17◦35.390 N, 24◦15.120 W) as well as the track of the cyclonic Back Close eddy (Argo2008, red line), surveyed with an Argo float are shown. The position (dots) and dates (label) of the first and last identification of the three eddies are given. The RV METEOR survey Full Screen / Esc of the CVOO2007 (grey star) is labelled accordingly. For reference, the average footprint (circle at CVOO) is given. Positions of the Argo float profiles surveyed inside (white circles) as well as Printer-friendly Version outside the cyclonic eddy radius (crosses) are shown. Interactive Discussion

Open ocean 350 dead-zones 300

CVOO2007 CVOO2010 J. Karstensen et al.

] 250 1 −

200 Title Page mol kg µ 150 Abstract Introduction oxygen [ 100 Conclusions References

50 Tables Figures

0 2007 2008 2009 2010 2011 2012 2013 2014 J I

J I Figure 2. Time series of DO from the CVOO site at 40 to 60 m depth (black line) and ﬁrst period at 140 m (grey line). The passage of the two anticyclonic-modewater eddies in February 2007 Back Close (CVOO2007) and February 2010 (CVOO2010) are labelled accordingly. The theoretical oxygen Full Screen / Esc saturation (red line) is shown as well as the 40 µmolkg−1 threshold reported in the literature. The period from 15 January to 15 March for each year is indicated by grey shaded area. Printer-friendly Version

Interactive Discussion

a b 2010 c 2010 Open ocean 01.Feb 15.Feb 01.Mar 01.Feb 15.Feb 01.Mar 0 37 0 0.5 dead-zones 250 25.3 0.4 50 26.1 50 25.3 26.1 200 0.3 J. Karstensen et al. 36.5 100 100 ] 0.2 26.3 −1 26.3 26.5 150 26.5 0.1 150 150

26.7 36 0 26.7 26.7 26.7 Title Page 100 et [m] Depth 200 [m] Depth 200

xgn[mlkg [µmol oxygen −0.1

250 250 −0.2 Abstract Introduction 35.5 40 26.9 26.9 −0.3 300 26.9 300 26.9 Conclusions References −0.4 0 15.J an 01.Feb 15.Feb 01.Mar 2010 350 35 350 −0.5 −150 −75 0 75 150 −150 −75 0 75 150 Tables Figures Distance [km] Distance [km]

Figure 3. (a) Time series of DO from the two sensors available at nominal 42 m (black line) J I and 170 m depth (blue line). For reference, the oxygen saturation (grey line) as well as the J I 40 µmolkg−1 threshold from the literature (broken line) are shown. Corresponding time series of (b) salinity (in PSS-78) and (c) meridional ﬂow (m s−1) in the upper 350 m as observed during Back Close the CVOO2010 passage. The black lines in (b) indicate the varying depth of the oxygen sensors shown in (a). Selected potential density anomaly surfaces are shown as white contours in (b Full Screen / Esc and c) for reference and the time series data was converted into distance assuming an eddy translation speed of 7 cm s−1. Printer-friendly Version

Interactive Discussion

200 26.7 36 26.7 Salinity

Depth [m] 26.7 35.8 250 J. Karstensen et al. 35.6 300 26.9 26.9 26.9 35.4 350 26.9 35.2 400 35 Title Page 01.Mar 01.May 01.Jul 01.Sep 01.Nov 01.Jan 01.Mar 01.May c 2008 2009| b 0 200 0 25 24 Abstract Introduction 2526.1 24 50 50 26.3 26.3 26.1 100 100 Conclusions References 26.5 26.5 150 ]

150 − 1 150 Tables Figures 26.7 26.5 200 200 26.7 100 Depth [m] depth (m) 250 26.7 250 Oxygen [ µ mol kg 26.9 J I 300 300 26.9 26.9 40 350 26.9 350 J I 20 400 10 400 0 0.05 0.1 0.15 0.2 01.Mar 01.May 01.Jul 01.Sep 01.Nov 01.Jan 01.Mar 01.May −1 Back Close 2008 2009| aOUR (µmol (kg day) )

Full Screen / Esc Figure 4. Time series of (a) salinity and (b) oxygen from profiling float data. The two neighbouring boxes indicate the period when the float was trapped in the cyclonic eddy Printer-friendly Version separating the isolated period (left box) and the non-isolated period (right box). Potential density anomaly contours are shown as contour lines. (c) Vertical profile of the aOUR derived from Interactive Discussion successive dives during the period when the eddy was isolated.

Open ocean dead-zones a b o o J. Karstensen et al. 22 N 0.6 22 N 0.6 25. Nov. −1. Dec. 2009 9. J an. −16. J an. 2010 0.5 0.5

o o 20 N 0.4 20 N 0.4 Title Page ) ) −3 −3

0.3 0.3 18oN 18oN Abstract Introduction

Chl−a(mgm Chl−a(mgm Conclusions References 16oN 0.2 16oN 0.2 Tables Figures 0.15 0.15 14oN 14oN 26oW 24oW 22oW 20oW 18oW 26oW 24oW 22oW 20oW 18oW J I Figure 5. Surface chlorophyll concentration of the CVOO2010 anticyclone at two life stages approximately two month (a) and one month (b) before the centre of the anticyclone crossed J I the CVOO mooring. The SLA derived track of the anticyclone centre (compare Fig.1) and the Back Close approximate diameter (130 km) are shown for reference. The white circle with a plus sign marks the CVOO position. Full Screen / Esc

Printer-friendly Version

Interactive Discussion

a b BGD 0 200 0 0.5

24 0.4 25 25 25 24 25 ) 11, 17391–17411, 2014 0.3 −1 50 26 150 50 26 26.1 26.1 0.2 75 26.3 75 0.1 26.3 Open ocean 100 100 100 0 Depth Depth 125 Oxygen 125 −0.1 dead-zones 26.5 26.5 −0.2 150 150 26.7 −0.3 40 (m s Velocity Merdidional J. Karstensen et al. 175 175 26.7 −0.4 20 200 10 200 −0.5 19°00’W18°30’W18°00’W17°30’W 19°00’W18°30’W18°00’W17°30’W

c Title Page 22 25 25.4 Abstract Introduction

20 25.4 Conclusions References 26 26.1 18 26.3 Tables Figures

C) 26.5 °

Θ ( 16 26 26.7 26.1 J I 26.9 26.3

14 26.5 J I

26.7 Back Close 12 26.9

35.2 35.4 35.6 35.8 36 36.2 36.4 36.6 36.8 Full Screen / Esc salinity

Printer-friendly Version Figure 6. Vertical distribution of (a) oxygen (µmolkg−1) and (b) meridional velocity (m s−1) ◦ ◦ surveyed with RV METEOR (cruise M68/3) on the 18 July 2006, at 18 N and from 17 to 19 W Interactive Discussion (see Fig. 1 for position). (c) Hydrographic characteristics of the eddy core as observed with RV METEOR (red lines) and as observed 7 month later during the CVOO2007 passage (blue dots). For reference typical background conditions at CVOO are shown (magenta dots and stars).

−1 Open ocean 200 dead-zones

100 J. Karstensen et al. xgn[mlkg [µmol oxygen

0 Title Page 01.Dec 01.J an 01.Feb 01.Mar 01.Apr 01.May 01.J un 01.J ul −55 2009 | 2010 21 Abstract Introduction 17 −60 Conclusions References 13 −65 Tables Figures 9

−70 Targetstrength

UTC Hour of day 5 J I 1 −75 J I 01.Dec 01.J an 01.Feb 01.Mar 01.Apr 01.May 01.J un 01.J ul 2009 | 2010 Back Close

Figure 7. Time series of (upper) oxygen at nominal 45 m depth and (lower) target strength Full Screen / Esc between 65 to 70 m depth calculated from the 300 kHz acoustic Doppler current proﬁler data at CVOO2010 plotted against hours of the day (in dB). Minimal target strength during all hours Printer-friendly Version of day is seen during the passage of the anticyclonic-modewater eddy between 8 and 25 February 2010. Interactive Discussion

17411